So maybe it is right that in short channels the drain current is lower.
Maybe in short channel, the drain current reaches its saturation faster, and therefore it doesnt keep growing with VDS.
I think we've agreed that the saturation current can be reached with a lower Vds for a shorter channel. BUt just because the saturation current can be reached with a lower Vds does not necessarily mean that the value of the saturation current is less than a longer channel. (A short channel certainly has increased current capability due to reduced heating from resistive losses, but I don't see why the saturating current would be less than a longer channel).
I'm not saying you're wrong, I'm just saying the Vds for current saturation and the amps of the saturating current are not necessarily related proportionally (at least from what I've been reviewing so far. They could be.). It's just from all the FETs I've looked at that come from the same series, the lower voltage FETs seem to have a higher saturation current (and higher current capability due to lower resistance) than an otherwise equivelant higher voltage FETs.
Saturation current is reached only when velocity saturation is reached?
I mean, if the channel is pinched off at the darin side, but velicity saturation is not reached, then we the current will keep growing with the rise of VDS?
I don't know this without reviewing some more. There might be more than one cause for current saturation, just like in diodes where there is more than one way for the diode to breakdown. It could very well be that people just say saturation to save the effort of saying velocity saturation all the time.
EDIT - FURTHER UPDATES:
Here is more reading that may be helpful:
The electrical engineering handbook - Google Book Search
The first page (right hand columns) confirm that the Vds required to reach saturation is less in short channels than long ones (woo hoo! I learned enough in school to relearn this stuff and surmise things about it!).
But the equation they give short afterward (equation 3.4) seems to imply that the value of the saturation current is dependent on channel WIDTH not length. So I'm starting to think that MOSFETs that had different channel lengths, but were otherwise identical would have the same saturation current, regardless of channel length (although Vds required for saturation would be different since that does depend on channel length). Earlier, I was thinking that this would be the case if velocity saturation IS saturation (the only effect causing saturation) but didn't say anything because I had no evidence that velocity saturation was the only effect causing saturation.